Track record
How the composite signals actually resolved. Hit rate is the share of closed signals that moved the predicted way — not a promise of returns.
Showing the current engine’s record — signals fired since 2026-07-04, when the balanced-book scorer went live.
By verdict
n= is shown on every bucket on purpose: 0% over five signals must not read like 0% over three hundred.
A signal counts as a hit here if it closed in profit over its hold — so the bar to beat is roughly 50%, the rate a coin flip would manage (roughly, because fees and the market’s own drift both nudge it). This is a different measurement from the stop/target record, which asks whether price reached the target before the stop and has a much lower chance bar.
Market cycle
A current read built only from data the engine already records — its own recent BUY/SELL mix and the Fear & Greed value it stamps on every signal. Every weighted input is listed under the dial.
Cumulative signal equity
indexed to 100Full history
daily closed & rolling hit rateDaily closed-signal count with the rolling 30-day hit rate, full history, with the 2026-07-04 engine-epoch change marked. Same outcome-audited source as the ledger.
Conviction & calibration
Does a higher stated conviction actually pay more? And when the model says 80%, do 80% land? The plain-English read leads; the reliability diagram underneath is the same data in its technical form.
Does higher conviction pay? Average move per signal
Reliability of stated confidence vs actual win-rate
Every resolved signal
none yetNo resolved signals in the window yet
Signals land here once they resolve (hit target/stop or age out). The live calibration proof above holds regardless.
Why this record can't be faked
Why this page exists.Most signal tools show win-rates you can’t verify. VYLOS Signal scores every signal it fires against the realized outcome and publishes the calibration here — if the model says 80%, this page shows whether ~80% actually win. It updates live, so it can’t be cherry-picked.
Independently verifiable.Every completed day’s signals are hashed (SHA-256) into a chain published with no login at /public/track-record/attestations, and a daily CI job commits the chain to the GitHub repo. Rewriting any historical call would break every later hash against the committed copy — the record can’t be quietly edited after the fact.
Anchored beyond our control. The chain tip is being submitted to public OpenTimestamps calendars, which commit it into Bitcoin. The first anchor lands within minutes of each deploy; once committed, proofs appear at /public/track-record/anchors and verify with the standard ots client — rewriting this history would then require rewriting Bitcoin, not just our server or our git repo.
This page reports a signal-resolution record, not actual trading results — no fees, slippage, or execution are modeled. Hypothetical and past results are not indicative of future returns.